Underground roadway point cloud boundary line extraction method based on deep learning

The invention discloses an underground roadway point cloud boundary line extraction method based on deep learning, and relates to the technical field of mine point cloud boundary line extraction. The method comprises the following steps: firstly, collecting roadway point cloud data through a laser radar, and then carrying out denoising, feature retention and feature reconstruction on the point cloud data; then data slicing is carried out, the anchor rods under the point cloud view angle are marked, and a training set is made; establishing a point cloud identification network, and identifying the anchor rod and a central point thereof; and finally, recording the position of the anchor rod in the center of the roadway wall according to the roadway slice and the identification result, and marking red. And connecting red marking points on the roadway slices, successfully drawing roadway boundaries, and forming a map and storing the map. According to the method, the point cloud data is acquired by using the laser